The future of population dynamic studies using marked individuals: A statistician's perspective

Capture-recapture methodology sensu late has grown since 1965 at a rate of 6.6% per year. Many generalizations have appeared, together with a progressive standardization in statistical approaches. This process of growth-generalization-standardization-has fairly close parallels in the history of tools used in human health studies. Three lines of development are considered in this review: the consolidation a recent results, the perspective of broad generalizations, and problems associated with transfer of Knowledge. A broader use of contingency table techniques and of likelihood ideas to increase robustness illustrates what may be expected from consolidation. In terms of generalizations, it seems we have entered an era of diversification, with great expectations in random effects models, individual covariate models and competing event models, all of which will make it possible to approach more successfully the biological questions dealing with individual variability. The transfer of knowledge has obviously been slower than in human health studies. One of the responsibilities of statisticians is to try to accelerate it.